The foundation of Theka: every document lives as a set of versions with a clear status, goes through its approval and signature workflow, is classified by category and type with the fields that really matter, and is linked to customers, suppliers and products.
In Theka a document is not a single file but the set of its versions. When you upload a new one, the previous one stays in the archive, tracked and downloadable. Each version has its own status in the lifecycle, so it is always clear which one is authoritative.
The final version is the one in use; signed ties it to the proof of signature; archived keeps it available but out of the everyday views. The status follows from the workflow: if a document type has no steps, the version is born final.
For each document type you define a sequence of steps: a simple go-ahead, an approval or an electronic signature. An approval can require the OK of everyone, of at least one person, or of people in order; the answers are approved, rejected or changes requested.
The category classifies (contracts, invoices, manuals, photos); the document type defines its structure: which fields are mandatory, which workflow it follows, which images it must have. No generic forms that are the same for everything.
Text, dates, numbers, choices, references to a customer or a product. Validation happens on the server; the AI can fill them in by reading the document.
Each document type can have its own “AI context”: a few lines that tell the AI what matters in that document. It guides summaries, fields, image descriptions and knowledge.
“Technical data sheet for an industrial pump: pressure, flow rate, materials and certifications matter. The product code is in the title.”
Everything that recurs across different documents becomes a master data record with its own fields and images. Documents reference it, search knows it, and the chat knows who you are talking about.
A document lives in several dimensions at once: you don't have to choose a single folder to put it in.
The main classification: contracts, invoices, manuals, technical data sheets, photos. Each category has its own fields, its own permissions and its own AI configuration.
They group documents from different categories around a job or an initiative, without moving them.
Free labels, more than one per document, filterable in search.
Incoming, outgoing or internal document; in multi-site companies, the site it belongs to.
From one file at a time to an entire archive: Theka processes in batches, shows an estimate and progress, and retries only what didn't go through.
Drag hundreds of files at once: Theka processes them in batches, with real-time progress.
Connect a SharePoint, S3 or SFTP folder: new or changed files come in on their own, deleted ones go to the bin.
The same file uploaded twice is recognised and flagged, instead of multiplying in the archive.
Your ERP or PIM uploads and reads documents with a dedicated key and its own permissions.
Need to send a document to someone who has no access to Theka? You create a link, even to a specific version, and revoke it whenever you like. Every opening and every download stays in the log.
Whoever receives the link sees exactly that version, not a copy that goes out of date.
The link closes whenever you want, or on its own when it expires.
Creation, opening, download: with date, time and origin.
The previous version stays in the archive, available to view and download; the new one becomes the current version and starts its approval workflow from the beginning.
Yes. The workflow is defined per document type: a contract can have review, approval and signature, an internal note no steps at all.
It can: customers, suppliers and products can be created by hand, imported, or brought in from the ERP through a connector or the API.
No. Versions, approvals, master data, search and signature all work with the AI features switched off.
A 30-minute call for a guided demo and an initial assessment of how well Theka fits the way you work.
